Enables humans and large language models to play Reconnaissance Blind Chess against AI bots or on the official remote server.
This tool serves as a Model Context Protocol (MCP) server, allowing both human players and LLMs to engage in Reconnaissance Blind Chess (RBC). It facilitates local games against a selection of AI bots (RandomBot, AttackerBot, TroutBot) and connects to remote games hosted on rbc.jhuapl.edu. The server integrates fully with MCP clients like Claude Desktop, providing comprehensive game state tracking and asynchronous execution to manage multiple concurrent games. It simplifies the complex turn structure of RBC, including sensing and moving phases, through a set of intuitive MCP tools.